survey panel data challenges

Measuring Cross-Channel Shoppers Across Platforms

Gian Fulgoni, Co-Founder & Chairman Emeritus at comScore began by noting that, while technology has had a tremendous impact, a negative consequence is that it is being used by people who can’t interpret relevant/usable data. Specifically, he cited the following problem with survey panel data: the cost to run panels has dropped and, as a result, panels are being used in negative ways with major repercussions.

He cited several examples where data has been misinterpreted:

  • A recent National Retail Federation survey that showed that Thanksgiving sales dropped 11% and that 42% of all buying had occurred online. As a result of this data, retailers’ stock prices dropped. Meanwhile data from the U.S. Department of commerce revealed that only 14% of Thanksgiving commerce was occurring online.
  • Mary Meeker’s 2015 Internet Trend Report found that, in 2014, people spent 5.0 hours per day online more time than the 4.3 hours they spend watching TV. Meanwhile comScore data reveals that online time is 70% of TV time.
  • Online surveys actually overstate time spent online by 35% and understate TV by 20%.

According to Gian, there are two main reasons as to why these types of discrepancies occur:

  • Online survey panelists are heavier-than-average Internet users.
  • Consumers are poor at recall.

Key takeaways

  • Be careful when using a survey to ask people to remember what they bought: “If you want to know how much money and time people spend online, don’t ask them”.
  • It’s good to include passive behavioral panels that don’t require consumer recall and integrate findings with census level data.

See all 5 Cups articles